Pasar al contenido principal

IA explicable para desinformación y detección de conspiración durante infodemias (XAI-DisInfodemics)

PLEC2021-007681

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Convocatoria Proyectos de I+D+i en Líneas Estratégicas
Año convocatoria 2021
Plan Estatal 2017-2020 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSITAT POLITÈCNICA DE VALÈNCIA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 12
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

What distinguishes conspiracy from critical narratives? A computational analysis of oppositional discourse

  • Korencic, Damir
  • Chulvi, Berta
  • Bonet Casals, Xavier
  • Taulé, Mariona
  • Toselli, Alejandro Héctor|||0000-0001-6955-9249
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] The current prevalence of conspiracy theories on the internet is a significant issue, tackled by many computational approaches. However, these approaches fail to recognize the relevance of distinguishing between texts which contain a conspiracy theory and texts which are simply critical and oppose mainstream narratives. Furthermore, little attention is usually paid to the role of inter-group conflict in oppositional narratives. We contribute by proposing a novel topic-agnostic annotation scheme that differentiates between conspiracies and critical texts, and that defines span-level categories of inter-group conflict. We also contribute with the multilingual XAI-DisInfodemics corpus (English and Spanish), which contains a high-quality annotation of Telegram messages related to COVID-19 (5000 messages per language). We also demonstrate the feasibility of an NLP-based automatization by performing a range of experiments that yield strong baseline solutions. Finally, we perform an analysis which demonstrates that the promotion of intergroup conflict and the presence of violence and anger are key aspects to distinguish between the two types of oppositional narratives, that is, conspiracy versus critical., This work was done in the framework of the research project XAI-DisInfodemics: eXplainable AI for disinformation and conspiracy detection during infodemics (PLEC2021-007681) funded by MICIU/AEI/10.13039/501100011033 and by 'European Union NextGenerationEU/PRTR'. We would like to thank Mariona Coll Ardanuy for all the helpful advice pertaining to the improvement of the article text and concept.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

The Impact of Psycholinguistic Patterns in Discriminating between Fake News Spreaders and Fact Checkers

  • Giachanou, Anastasia
  • Ghanem, Bilal Hisham Hasan
  • Rissola, Esteban A.
  • Crestani, Fabio
  • Oberski, Daniel
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] Fake news is a threat to society. A huge amount of fake news is posted every day on social networks which is read, believed and sometimes shared by a number of users. On the other hand, with the aim to raise awareness, some users share posts that debunk fake news by using information from fact-checking websites. In this paper, we are interested in exploring the role of various psycholinguistic characteristics in differentiating between users that tend to share fake news and users that tend to debunk them. Psycholinguistic characteristics represent the different linguistic information that can be used to profile users and can be extracted or inferred from users¿ posts. We present the CheckerOrSpreader model that uses a Convolution Neural Network (CNN) to differentiate between spreaders and checkers of fake news. The experimental results showed that CheckerOrSpreader is effective in classifying a user as a potential spreader or checker. Our analysis showed that checkers tend to use more positive language and a higher number of terms that show causality compared to spreaders who tend to use a higher amount of informal language, including slang and swear words., The works of Anastasia Giachanou and Daniel Oberski were funded by the Dutch Research Council (grant VI.Vidi.195.152). The work of Paolo Rosso was in the framework of the XAI-DisInfodemics project on eXplainable AI for disinformation and conspiracy detection during infodemics (PLEC2021-007681), funded by the Spanish Ministry of Science and Innovation, as well as IBERIFIER, the Iberian Digital Media Research and Fact-Checking Hub funded by the European Digital Media Observatory (2020-EU-IA0252).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Studying Fake News Spreading, Polarisation Dynamics, and Manipulation by Bots: a Tale of Networks and Language

  • Ruffo, Giancarlo
  • Semeraro, Alfonso
  • Giachanou, Anastasia
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] With the explosive growth of online social media, the ancient problem of information disorders interfering with news diffusion has surfaced with a renewed intensity threatening our democracies, public health, and news outlets¿ credibility. Therefore, thousands of scientific papers have been published in a relatively short period, making researchers of different disciplines struggle with an information overload problem. The aim of this survey is threefold: (1) we present the results of a network-based analysis of the existing multidisciplinary literature to support the search for relevant trends and central publications; (2) we describe the main results and necessary background to attack the problem under a computational perspective; (3) we review selected contributions using network science as a unifying framework and computational linguistics as the tool to make sense of the shared content. Despite scholars working on computational linguistics and networks traditionally belong to different scientific communities, we expect that those interested in the area of fake news should be aware of crucial aspects of both disciplines., The work of Paolo Rosso was carried out in the framework of the IBERIFIER hub on Iberian media research and factchecking funded by the European Digital Media Observatory (INEA/CEF/ICT/A202072381931 n.2020-EU-IA-0252), and the following projects: XAI-Disinfodemics on eXplainable AI for disinformation and conspiracy detection during infodemics (PLEC2021-007681) and MARTINI on Malicious Actors pRofiling and deTection In online social Networks through artificial Intelligence (PCI2022-134990-2) of the CHISTERA IV Cofund 2021 program, funded by MCIN/AEI/10.13039/501100011033, Spain and by the European Union NextGenerationEU/PRTR, Spain .
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

SceneFND: Multimodal Fake News Detection by Modeling Scene Context Information

  • Zhang, Guobiao
  • Giachanou, Anastasia
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] Fake news is a threat for the society and can create a lot of confusion to people regarding what is true and what not. Fake news usu ally contain manipulated content, such as text or images that attract the interest of the readers with the aim to convince them on their truthfulness. In this article, we propose SceneFND (Scene Fake News Detection), a system that combines textual, contextual scene and visual representation to address the problem of multimodal fake news detection. The textual representation is based on word embeddings that are passed into a bidirectional long short-term memory network. Both the contextual scene and the visual representations are based on the images contained in the news post. The place, weather and season scenes are extracted from the image. Our statistical analysis on the scenes showed that there are statistically significant differences regarding their frequency in fake and real news. In addition, our experimental results on two real world datasets show that the integration of the contextual scenes is effective for fake news detection. In particular, SceneFND improved the performance of the textual baseline by 3.48% in PolitiFact and by 3.32% in GossipCop datasets. Finally, we show the suitability of the scene information for the task and present some examples to explain its effectiveness in capturing the relevance between images and text., The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: The work of Anastasia Giachanou is funded by the Dutch Research Council (grant VI.Vidi.195.152). The work of Paolo Rosso was in the framework of the Iberian Digital Media Research and Fact-Checking Hub (IBERIFIER) funded by the European Digital Media Observatory (2020-EU-IA0252), and of the XAI-DisInfodemics research project on eXplainable AI for disinformation and conspiracy detection during infodemics, funded by the Spanish Ministry of Science and Innovation (PLEC2021-007681).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Automatic detection of health misinformation: a systematic review

  • Baris-Schlicht, Ipek
  • Fernandez, Eugenia
  • Chulvi-Ferriols, María Alberta
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] The spread of health misinformation has the potential to cause serious harm to public health, from leading to vaccine hesitancy to adoption of unproven disease treatments. In addition, it could have other effects on society such as an increase in hate speech towards ethnic groups or medical experts. To counteract the sheer amount of misinformation, there is a need to use automatic detection methods. In this paper we conduct a systematic review of the computer science literature exploring text mining techniques and machine learning methods to detect health misinformation. To organize the reviewed papers, we propose a taxonomy, examine publicly available datasets, and conduct a content-based analysis to investigate analogies and differences among Covid-19 datasets and datasets related to other health domains. Finally, we describe open challenges and conclude with future directions., This work was done in the framework of the XAI-DisInfodemics research project on eXplainable AI for disinformation and conspiracy detection during infodemics (Grant PLEC2021-007681 funded by MCIN/AEI/10-.13039/501100011033 and by European Union NextGenerationEU/PRTR) and the IBERIFIER on Iberian Digital Media Research and Fact-Checking Hub (supported by European Commission CEF-TC-2020-2, European Digital Media Observatory, under reference 2020-EU-IA-0252).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Overview of AuTexTification at IberLEF 2023: Detection and Attribution of Machine-Generated Text in Multiple Domains, Resumen de AuTexTification en IberLEF 2023: Detección y Atribución de Texto Generado Automáticamente en Múltiples Dominios

  • Sarvazyan, Areg Mikael
  • González, José Ángel
  • Franco-Salvador, Marc
  • Rangel, Francisco
  • Chulvi-Ferriols, María Alberta
  • Rosso, Paolo|||0000-0002-8922-1242
[EN] This paper presents the overview of the AuTexTification shared task as part of the IberLEF 2023 Workshop in Iberian Languages Evaluation Forum, within the framework of the SEPLN 2023 conference. AuTexTification consists of two subtasks: for Subtask 1, participants had to determine whether a text is human-authored or has been generated by a large language model. For Subtask 2, participants had to attribute a machine-generated text to one of six different text generation models. Our AuTexTification 2023 dataset contains more than 160.000 texts across two languages (English and Spanish) and five domains (tweets, reviews, news, legal, and how-to articles). A total of 114 teams signed up to participate, of which 36 sent 175 runs, and 20 of them sent their working notes. In this overview, we present the AuTexTification dataset and task, the submitted participating systems, and the results., [ES] Este artículo presenta un resumen de la tarea AuTexTification como parte del workshop IberLEF 2023 sobre el Iberian Languages Evaluation Forum, en el marco de la conferencia SEPLN 2023. AuTexTification consta de dos subtareas: en la Subtarea 1, los participantes tuvieron que determinar si un texto fue escrito por un humano o generado por un modelo de lenguaje masivo. Para la Subtarea 2, los participantes debían atribuir un texto generado automáticamente a uno de seis modelos de generación de texto diferentes. El conjunto de datos AuTexTification contiene más de 160.000 textos en dos idiomas (inglés y español) y cinco dominios (tweets, reseñas, noticias, legislación y artículos instructivos). Un total de 114 equipos se inscribieron para participar, de los cuales 36 enviaron 175 resultados y 20 de ellos enviaron artículos. En este artículo, presentamos el conjunto de datos y la tarea AuTexTification, los sistemas enviados por los participantes y sus resultados., The work from Symanto has been partially funded by the Pro<SUP>2</SUP>Haters - Proactive Profiling of Hate Speech Spreaders (CDTi IDI-20210776), the XAI-DisInfodemics: eXplainable AI for disinformation and conspiracy detection during infodemics (MICIN PLEC2021-007681), the OBULEX - OBservatorio del Uso de Lenguage sEXista en la red (IVACE IMINOD/2022/106), and the ANDHI - ANomalous Diffusion of Harmful Information (CPP2021-008994) R&D grants. The work of Areg Mikael Sarvazyan has been partially developed with the support of valgrAI - Valencian Graduate School and Research Network of Artificial Intelligence and the Generalitat Valenciana, and co-founded by the European Union. The research at the Universitat Politecnica de Valencia was framed under the FairTransNLP research project, Grant PID2021-124361OB-C31 funded by MCIN/AEI/10.13039/501100011033 and by ERDF, EU A way of making Europe.
RUA. Repositorio Institucional de la Universidad de Alicante
Publicación Otros (Other)

Overview of DETESTS at IberLEF 2022: DETEction and classification of racial STereotypes in Spanish, Resumen de la tarea de DETESTS en IberLEF 2022: DETEcción y clasificación de eSTereotipos raciales en eSpañol

  • Ariza, Alejandro
  • Schmeisser-Nieto, Wolfgang S.
  • Nofre, Montserrat
  • Taulé Delor, Mariona
  • Amigó Cabrera, Enrique
  • Chulvi, Berta
  • Rosso, Paolo
This paper presents an overview of the DETESTS shared task as part of the IberLEF 2022 Workshop on Iberian Languages Evaluation Forum, within the framework of the SEPLN 2022 conference. We proposed two hierarchical subtasks: For subtask 1, participants had to determine the presence of stereotypes in sentences. For subtask 2, participants had to classify the sentences labeled with stereotypes into ten categoriesEste artículo presenta un resumen de la tarea DETESTS como parte del workshop IberLEF 2022, dentro de la conferencia SEPLN 2022. Proponemos dos subtareas jerárquicas: En la subtarea 1, los participantes tuvieron que determinar la presencia de estereotipos raciales en oraciones. En la subtarea 2, de las oraciones etiquetadas con estereotipo, los participantes tuvieron que clasificarlas en una o más de diez categorías. El dataset DETESTS contiene 5.629 oraciones de comentarios que responden a artículos de periódicos sobre inmigración en español. 51 equipos se registraron para participar, de los cuales 39 enviaron predicciones de sistemas y 5 de ellos enviaron artículos. En este artículo presentamos información sobre los datasets de entrenamiento y de prueba, los sistemas utilizados por los participantes, las métricas de evaluación y sus resultados.. The DETESTS dataset contains 5,629 sentences in comments in response to newspaper articles related to immigration in Spanish. 51 teams signed up to participate, of which 39 sent runs, and 5 of them sent their working notes. In this paper, we provide information about the training and test datasets, the systems used by the participants, the evaluation metrics of the systems and their results.
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Otros (Other)

On the Design of a Misinformation Widget (MsW) Against Cloaked Science

  • Arroyo Guardeño, David
  • Degli Esposti, Sara
  • Gómez Espés, Alberto
  • Palmero Muñoz, Santiago
  • Pérez Miguel, Luis
Amongst all types of fabricated information travelling on open social networks, scientific disinformation, or cloaked science, is both insidious and challenging to be investigated. Here we present the design of the TRESCA misinformation widget (MsW), which is both a methodology and a toolbox for investigating disinformation operations leveraging scientific communications. In developing MsW we adopt a human-in-charge approach to AI: the automated tools included in MsW REST API are meant to support, not to substitute or undermine, users’ decision-making capacity. On the journey toward information verification, MsW AI toolbox helps users test both the veracity of claims and the reliability of sources. While the toolbox integrates open source intelligence solutions, MsW methodology fosters users’ critical thinking., This work was partially funded by European H2020 project TRESCA (Grant Agreement No 872855), national project XAI-DisInfodemics (grant PLEC2021-007681 funded by MCIN/AEI/10.13039/501100011033 and by European Union NextGeneration EU/PRTR), and regional project “CYNAMON - Cybersecurity, Network Analysis and Monitoring for the Next Generation Internet” (funded by “Programas de Actividades de I+D entre grupos de investigación de la Comunidad de Madrid en tecnologías 2018” P2018/TCS-4566; BOCM. No. 304; 21/12/2018)., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

A cybersecurity risk analysis framework for systems with artificial intelligence components

  • Camacho Sosa-Días, Jorge
  • Couce-Vieira, A.
  • Arroyo Guardeño, David
  • Ríos Insua, David
28 páginas, 6 figuras, 9 tablas, The introduction of the European Union Artificial Intelligence (AI) Act, the NIST AI Risk Management Framework, and related international norms and policy documents demand a better understanding and implementation of novel risk analysis issues when facing systems with AI components: dealing with new AI- related impacts; incorporating AI-based assets within the cyber architecture; considering AI-based security and recovery controls within the cybersecurity portfolio; and managing novel AI-based targeted attacks. This paper suggests solutions to such issues and integrates them within a broad novel framework to support risk analysis in systems with AI components and services. An example concerning automated driving systems illustrates the framework validating it conceptually., The work was supported by the AXA-ICMAT Chair in Adversarial Risk Analysis, the EU’s Horizon 2020 Research and Innovation Programme under Grant Agreement No. 101021797 (STARLIGHT), the Spanish Ministry of Science program PID2021-124662OB-I00, AFOSR Eu- ropean Office of Aerospace Research and Development award FA8655-21-1-7042, a fellowship from “la Caixa” Foundation (ID 100010434), LCF/BQ/DI21/11860063, and grant PLEC2021- 007681 funded by MCIN/AEI/ 10.13039/501100011033, and by European Union NextGenera- tion EU/PRTR, Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

WIDISBOT: Widget to analyse disinformation and content spread by bots.

  • Camacho, José Manuel
  • Pérez Miguel, Luis
  • Arroyo Guardeño, David
6 páginas, 3 figuras, The increasing prevalence of bots poses a sig- nificant challenge for maintaining the integrity of online information. Bot campaigns have been deployed for both economic scams and political interference, making it necessary to develop a system to detect these agents and ana- lyze their behavior. We present a scalable appli- cation designed to identify bots and to buttress the investigation of disinformation campaigns. Our intention is to provide professionals with- out technical expertise with an effective tool to identify and analyze content generated by bots. This will enable researchers from diverse backgrounds to study bot activity, fostering an interdisciplinary understanding of the strategies these agents use to spread disinformation, and the characteristics of their discourse. We illus- trate how to use the application through a case study on COVID-19., This work was supported by the Spanish Ministry of Science program PID2021-124662OB-I00; a fellowship from ”la Caixa” Foundation (ID 100010434), whose code is LCF/BQ/ DI21/11860063; and Grant PLEC2021-007681 (project XAI-DisInfodemics) funded by MCIN/AEI/ 10.13039/501100011033 and by European Union NextGeneration EU/PRTR; and BBVA Foundation project AMALFI., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Contribución de congreso (ConferenceOutput)

WIDISBOT. Widget to analyze disinformation and content spread by bots.

  • Camacho, José Manuel
  • Pérez Miguel, Luis
  • Arroyo Guardeño, David
Proceedings of the 4th Conference on Language, Data and Knowledge, 12 september 2023, Vienna, Austria. 6 páginas, 3 figuras, 1 tabla, The increasing prevalence of bots poses a sig- nificant challenge for maintaining the integrity of online information. Bot campaigns have been deployed for both economic scams and political interference, making it necessary to develop a system to detect these agents and ana- lyze their behavior. We present a scalable appli- cation designed to identify bots and to buttress the investigation of disinformation campaigns. Our intention is to provide professionals with- out technical expertise with an effective tool to identify and analyze content generated by bots. This will enable researchers from diverse backgrounds to study bot activity, fostering an interdisciplinary understanding of the strategies these agents use to spread disinformation, and the characteristics of their discourse. We illus- trate how to use the application through a case study on COVID-19., This work was supported by the Spanish Ministry of Science program PID2021-124662OB-I00; a fellowship from ”la Caixa” Foundation (ID 100010434), whose code is LCF/BQ/ DI21/11860063; and Grant PLEC2021-007681 (project XAI-DisInfodemics) funded by MCIN/AEI/ 10.13039/501100011033 and by European Union NextGeneration EU/PRTR; and BBVA Foundation project AMALFI., Peer reviewed
Archivo Digital UPM
Publicación Artículo científico (JournalArticle)

Exploring Dimensionality Reduction Techniques in Multilingual Transformers

  • Huertas Tato, Javier|||0000-0003-4127-5505
  • Huertas García, Álvaro|||0000-0003-2165-0144
  • Martín García, Alejandro|||0000-0002-0800-7632
  • Camacho Fernández, David|||0000-0002-5051-3475